За это сейчас воюют Китай и США

In your pocket lies the control panel for the

most dangerous weapon

ever created by mankind. And

right now China, the US and Europe

are fighting over who will own it.

This war is not fought on the battlefield, but in

deep mines where metals and

minerals are mined, as well as in scientific laboratories

and the offices of the planet's leading IT corporations.

They put all their money into making sure that

this weapon belongs to

them [music]. While we're asking Iya

what the weather's like today, the government is

using neural networks to plan entire

military operations, like kidnapping

Nicolas Maduro or attacking Iran. And it’s

not for nothing that the richest man on the planet, Elon Musk, says the following things. I

think the danger from AI is much greater

than the danger from nuclear warheads. But

creating a neural network is not just like

coming up with a program for a smartphone. This

means controlling the largest

production chain in the history of the global

economy. I'll tell you how just a

few companies are holding the entire

AI industry hostage,

and which countries are already winning the

neural network war that has been going on since

2014. You will see which places on the planet

have become the hottest spots of this

war, while you, without suspecting anything,

ask the neurons your most intimate

questions. After watching this video, you'll

look at artificial intelligence differently

because you'll learn how we feed

the most dangerous monster on the planet

every day and can't stop.

My name is Nikolai Myachin, and this is Simple

Economics. Let's figure out how we will

destroy ourselves.

[music] On

February 28, 26, the United States and

Israel

killed Iran's Supreme Leader Ali

Khamenei in his Tehran residence with a single, precise missile strike. The

American neural network Claude

developed this operation in 90 minutes. Not hundreds of

analysts locked in for

months, but just a couple of the right

queries to artificial intelligence. How is

this possible magic? When you write in

GPT chat about how to get a mortgage at 5%, it doesn't

understand you as a person. It breaks down

the query into small parts and then

selects the most appropriate answer

based on the combination of words and context. That

is, the neural network does not think in the usual

sense, but calculates the most

appropriate answer. And the better it is

trained on huge amounts of data, the more

accurate this answer will be. But remember

how your child, younger brother or

nephew grew up. At first, all he could do was

lie in his crib and coo. But at

six months the baby learned to sit, at one year he

walked, at two he talked, and at 7:00

he went to first grade. [music] So do

neural networks. Every time you

ask her something and give her information about

yourself, you feed her new data. And

when enough is new, the

developers release a new model

that knows even more and gives even

more accurate answers. So, the neural network

turns from a baby into the wisest

advisor on the planet. The experience of a neural network

model is stored in its weights. It is such a

set of numerical parameters, well, like a

kind of DNA. And this DNA is the most

deadly weapon. If you think

I've gone crazy and joined the sect of

accusers of all mortal sins,

then look at what American officials are writing

. Direct access to model

scales will allow the neural network to be used

to create chemical or

biological weapons. And the risk for now is not

that the neural network will one day wake up and

destroy humanity, but that

the neural network might end up in the wrong

hands. Not long ago, on June 11,

2026, a model from the

Anthropopic company hacked

the servers of the

US National Security Agency in a matter of hours and gained access to

internal files. What had been built up over

the years as an impenetrable defense, the neronka

destroyed in a matter of hours. This is why

countries close their neural networks to

others. When Antropic

released its new popular

neural network, the Fable [music] 5, on June 9th, just

three days later the US government forced

the company to restrict access to it for

foreigners because the Fable

5 can even hack corporate

security systems. Anthroped had to

close the model and urgently make changes to it

to make it safe. There is already a shortage of

highly trained cybersecurity specialists who can identify

vulnerabilities in security systems and help

fix them.

Therefore, to gain practical skills for

work in this field, I recommend the

online master's program in information

security from Skill Factory and MIFI,

as their department of cryptology and

cybersecurity is among the

top three departments in Russia for training

specialists in computer

security.

You can apply all the knowledge you gain on

licensed equipment directly in

ME laboratories thanks to a license from the

Federal Service for Technical and

Export Control. Here you will receive a

systematic foundation from the first weeks of

training. You will learn to speak the same

language as business. You will master the full

security cycle, from infrastructure code to

business processes. You will gain a unified

understanding of security, from cryptography and

operating system protection to legal

frameworks and security architecture.

Online Master's degree is a modern

format of education. It is in-person and provides

full access to all student

benefits, such as military service deferments and

academic leave, but is conducted

entirely online. And there is also

government support. The cost of an

educational loan starts from 225 rubles. per

month for the first semester. Classes are taught

not only by MEPhI's best professors, but

also by Cybersecurity staff from Sber,

Kaspersky Lab, CryptoPro, and

other companies for whom

information security is a

natural working environment.

Learning from scratch is the key to any

completed higher education. To

prepare for entrance examinations,

Skill Factory offers a free

preparatory course and

consultations. By submitting your application

now, you can gain access to

additional events and materials from the

OTFI even before the course begins. So,

to gain strong

practical skills in Cybersecurity,

follow the link in the description or

this code on the screen. Admission to the program is

now open, and places are limited,

so hurry.

And here's the most paradoxical thing: today, no

one will ban or destroy

artificial intelligence, because people have

already become accustomed to neural networks doing routine, and sometimes even the

main work, for them,

while businesses have begun to implement neural networks in

every nook and cranny to squeeze out maximum

profits, laying off thousands of now

useless workers. For example,

in Indian factories, workers

are wearing cameras so that artificial

intelligence can then train robots to do this work

, which will replace people. And in

New York today you will get paid to

clean your apartment. Everything to

record this process on camera and then

train cleaning robots.

But the state is not concerned about future

unemployment now, but rather about the fact that if you don’t want

to be destroyed, you need to create

your own models. Today, the United States is ahead of the entire

planet at the first level and chain

. They created the most

popular neurocities in the world: Cha GPT,

Demminay, Claude and, of course, Grock.

Each of these models is tailored to its own

specific task. This is how open AI is trying to

turn GPT chat into a universal

assistant for all occasions. Google

is also turning Jemin into an assistant

that integrates into all of its services. And

given that Google is used by nine

out of ten internet users,

the company has billions of customers

who make Google smarter every day

. Grok is spying on the live

feed of a social network banned in Russia

. So, AI knows what's

being discussed right now, but it's the perfect

tool for finding and analyzing the

latest information and the future foundation

for the most sophisticated propaganda. You might

think that the creation centers are the

IT company headquarters in

San Francisco and London, but this is not the case.

Behind this first level,

where developers sit and write code,

there is a real iceberg of

bottlenecks in which the

entire global industry is crowded. It is precisely on these very

bottlenecks that the outcome of the

geopolitical pursuit to create

humanity's most dangerous weapon,

which is happening right now before our

eyes, depends. To understand economics, you

only need to start with one book.

More than 10,000 copies sold.

Average rating 4.9. Book Simple

Economics. Search on marketplaces or by the

link [music] in the description.

When you message GPT chat about how to get

a mortgage at 5%, the signal from your

smartphone or laptop enters the

internet and travels through fiber-optic

cables to the load balancer. This is a

control room that evaluates

server availability, load,

latency, user region, and

hundreds of other parameters so you get a

response in 5 seconds, not a minute.

And after selecting a server, the balancer

sends your request there. And then

the magic, or rather the physics of computing,

happens in the humming data centers

scattered around the world. Here's what

Google's Alabama data center looks like,

where

power disappears like a black

hole. In general, all the data centers that

exist on the planet today

consume more electricity in a year than, for

example, the whole of France. Electricity

powers the heart of artificial intelligence.

Millions of graphics

processors like these. It is the processor, or rather

its brain, a miniature chip, that does all the

work. It selects the

most appropriate answers to your requests to

tell you how to get a cheap

mortgage. Several processors

are combined into one cluster to later

create an entire supercomputer in this

data center. Why is this necessary? Ah, well,

imagine that your question to the GPT chat

is a sack of potatoes. You can take a knife

and start peeling potatoes one by one, but

if GPUs worked like

that, then the answer to a simple morning

question: what should I wear today? You would

receive it already in the evening before going to bed. So the

GPU divides a sack of

potatoes between hundreds of assistants

who peel them simultaneously, and you

get the answer in 5 seconds. While

GPUs are ideal

for running artificial intelligence,

they were originally designed for

video game graphics. Yes, it's all because of

those who already pre-ordered GTA 6 and took a

vacation until the end of November 2026

. It is because of them that we will be destroyed by

artificial intelligence. To avoid

growing and become smarter, they

need more new data centers. But it

won’t be possible to build up the entire land with these hangars. And here we

encounter the first bottleneck

. Firstly, electricity is not

infinite. Before building a new

data center, a power source must be found somewhere

, otherwise the power will go out in millions of apartments

nearby. And secondly,

there's no point in building a data center if you

can't buy as many chips as

you need. Well, imagine this: one

of the most in-demand

technologies today is in terrible

shortage, and chip manufacturers would be happy

to make more of them, but they themselves are running

into a shortage of components. A chip is a

silicon crystal containing thousands of

computing cores, which is the brain of

artificial intelligence. The chips are

so small that they are literally

drawn layer by layer onto the surface of

silicon wafers. This process

is called photolithography.

The plate is first polished to a mirror

finish and then coated with a special

light-sensitive material. After

this, a template with a drawing of the

future chip is placed and sent to a special

EUV lithograph. Inside this machine,

tiny drops of molten

tin fly. Powerful lasers hit them and the tin

turns into plasma. Plasma emits

extreme ultraviolet light.

This light is then collected by a system of

mirrors, directed through a special

template, and transfers

the outlines of the future chip elements onto a silicon wafer. This

process is repeated many times

because the chip is not a single pattern. It is a

multi-story microscopic city where

transistors are formed on the lower level. These are miniature

electric current switches. Above

them are insulating layers, metal

tracks, and so on. And each layer

must fit perfectly on the previous one.

If the pattern shifts even a nanometer,

everything will go to waste. And only one company in the world can produce such complex

equipment, the size of a school

bus weighing 180 grams, where

miniature chips are printed with microscopic

precision.

It seems that this is where this whole

iceberg and industry begins. [music]

But no, its roots go kilometers

underground to where quartz is mined.

In general, quartz is the second most

common metal on the planet.

But usually it contains a lot of impurities:

iron, aluminum, potassium. But for the

lithograph to work its magic, the

flint plate, which

is made from quartz, must be

crystal clear. The impurity content must

exceed 10 parts per million. In

percentage terms, this number looks like this: 99% pure

quartz is 999.

And there is only one place on the entire planet

where

such pure quartz can be mined on an industrial scale.

Sprucepine.

380 million years ago,

two lithospheric plates,

Africa and North America, collided in this area. The rock

became extremely hot after the accident, and since it

happened at a great depth, water

could not penetrate there and bring with it

various congestion. Eventually, the quartz

solidified into crystal clear material, making it

possible to produce chips for our

smartphones and computers. [music]

So it turns out that the modern chip

is a miniature microcircuit

that has taken literally the entire

global economy hostage. Without this

fingernail-sized component, a car wouldn't move or an

airplane fly, and there's simply no replacement for this entire

chain, as it's the

most efficient technology for

processing big data today. But that's

not all. The second key element, without

which the neural network will not work,

is ultra-fast memory, soldered right

next to the crystal. It enables

thousands of cores to operate smoothly, without

allowing them to sit idle. This is why Ii

can answer your questions so quickly

. And until it became such a big part of

our lives, chip and memory manufacturers

remained in the shadows. But as soon as

neural networks took off, everyone suddenly

needed both chips and ultra-fast

memory. And it is clear that it is

simply physically impossible to increase production in the shortest possible time

. To do this, we need to find

money, build new factories, hire and

train hundreds, if not thousands, of people.

This is how the whole industry looks in cross-section

. At the first level are the

developers of neural networks, Open AI, Google,

or Hiring a team of programmers and

writing code is perhaps the

easiest task in this entire chain,

which is why it is the biggest nesting doll. The

second level is data centers and a

smaller matryoshka doll. The most prominent

representative of this level is the

world's most valuable company, Nvidia, which

sells chips and other equipment for

data centers and earns

$200 billion a year from this. Of course, other

hardware manufacturers want to push

NVIDIA out of such a rich market. AMD,

Brotko, and even Google have joined the

race. And note that all of these companies are

entirely American. Therefore, the US has

every opportunity to tie their hands and

prohibit them from working with the wrong

countries. On January 15,

2025, the US Department of Commerce imposed

restrictions on the export of the most

advanced neural network chips from the country. All

other states were divided into three

groups. US allies such as Japan,

South Korea and the UK receive

advanced equipment without restrictions.

Around 120 countries, including Singapore, Israel, and the UAE,

can only rely on the supply of

previous-generation chips to

stay one step behind. But

Russia, China and Iran have no chance at all

here. Limit the key

component of future weapons creation, and

your opponents will be toothless.

China responded by developing its own

chips. And even after

Donald Trump visited

China on May 15, 2020 and allowed the Chinese to buy

previous-generation chips from the US, he did not

order a single such chip from the

Americans. Why? Because if

your country gets hooked on equipment

produced by your rival, it will

gain a powerful trump card in resolving any

disputes. So for China,

import substitution of American chips

is a matter of principle and sovereignty.

In general, there is nothing wrong with national chips or even

messengers. The

only question is how to use them.

For example, deliberately slowing down or

blocking competitors, but this is clearly not

ok. I explained what Max can expect with this approach

in my

Telegram channel, Simple Economy.

Read this post by following the link in the description

or by clicking on the code on the screen. There are already

more than 280,000 of us, and

we really understand what is

happening in Russia and the world and what we should

do about it. So, if you also

want to understand, then you are welcome.

Nvidia's products

have been replaced in the Chinese market by Huawei, which

has been on the US blacklist since

May 16, 2019. While

Nvidia's share of chips in China was

40% in 2025, it will drop to

8% in 2026, while Huawei's share of its own chips

will exceed 50%. At the same time,

Huawei's chips are still inferior to

Nvidia's most advanced developments, but they are superior to

their previous models, which Trump

kindly allowed to be purchased. Well, that's

not necessary anymore. Thank you. So, we

can safely add

another second-level flag to this map for

Huawei, a company from the Chinese city of Shenzhen.

It turns out that any sanctions are a

double-edged sword. On the one

hand, you limit your

opponent and get ahead while he is

dragging himself around somewhere. On the other hand,

there is always the risk that he will do something of

his own, yes, better, and not just

catch up with you, but overtake you, and you will

not be able to do anything about it. The leverage will be

lost, and in addition, your own

producer will become weaker, having lost

such a fat market. Incidentally, this is what

Nvidia's CEO himself says, and he's

certainly not thrilled about the

restrictions that are preventing him from

earning billions. In short, this is the

kind of struggle that is taking place at the second level and

chain, where the US is also in the lead, but

China is already hot on its heels. But

only the world's most valuable company,

Nvidia, doesn't manufacture processors and

chips from scratch. She assembles them from

ready-made components that are manufactured for her by

others. So now

NVIDIA is stuck in a traffic jam, waiting for deliveries of

chips and high-speed memory. Two

leading manufacturers of precisely this type of

memory, necessary for chips,

are located in one country in the world, South

Korea. These are SK H K Heinex with a share of 61% and

Samsung 17%. Between them is the American

Mikron with a share of 21%, meaning that the total for the

three of them is almost 100%. In the list of the most valuable

companies in the world, these three occupy

twelfth, thirteenth and fourteenth

places, respectively. Moreover, while

Samsung produces not only memory, but

also much more, such as smartphones and televisions,

SK HX is focused solely on

producing this ultra-fast memory.

Paradoxically, 20 years ago this

company almost went bankrupt

because this memory was not in demand at the time

. Well, today the same

product has made it the most valuable company in

South Korea. even allowed it to overtake

Samsung, which had been the leader for a quarter of a

century. All three say that

RAM supplies for this year,

2026, are already fully

booked, and the shortage is unlikely to disappear

even in 2027, although the

companies are actively building new factories.

Incidentally, it's not

just Nvidia that's lining up for memory, but also China's Huawei,

which could more quickly replace the

Americans in its market. But it

lacks components for chip production. But while two

South Korean companies share 100% of the market,

just one Taiwanese company has captured

90% of the

advanced chip market. This is TSMC. Well, here it is

in seventh place among the most expensive

companies on the planet. It turns out that the third

level and chains are controlled by

key suppliers of components for

the creation of graphics processors

that literally fit on one

hand. The first is Taiwan's TSMC, and the second,

third and fourth are SK, Kinix,

Micron and Samsung. These companies literally

hold up the entire global economy. Without their

chips and high-speed memory, the

production of smartphones, computers,

cars, bank cards,

telecommunications equipment, and,

of course, those very same data centers for

artificial intelligence would collapse. But if

this third matryoshka is just a few

companies, then this fourth one is just

one company, ASML,

from the Netherlands, where they manufacture those

very lithographs for printing chips. Yes,

lithographs are not only made in

the Netherlands. For example, the Japanese Nikon

produces DUV lithographs. This is the previous

generation of such machines. The difference between DUV

and EUV is the wavelength of ultraviolet light

they work with. And in Ivy it is 14

times shorter and is only 13.5 nm.

This is 6,000 times thinner than a human

hair. The shorter the ultraviolet wavelength,

the more accurately the chip can be drawn. Let's

take two pencils. One is sharply

sharpened, and the other has a thick rod.

And you will immediately notice the difference here. The same is true

for lithographers. The more accurate the drawing,

the more compactly

the transistors can be packed and a greater number of them can be placed

. This means the chip will become more powerful. And so far,

only the Dutch have been able to create such a

thin pencil, such a complex

machine that can draw chips with

precision down to the nanometer. In

2026, the company plans to

make and sell 60 machines, and in

2027, a whopping 80. Such a device

consists of, well, about 100,000

parts, many of which, yes, ASML

also orders from others. Well, for example, the

Carlisle company makes mirrors for reflecting ultraviolet radiation. And they have their own

suppliers. So this chain goes on and on

. However, the problem is not only

finding money and waiting in

line for an EUV lithograph. The supply of

such equipment requires an export

license from the Dutch government,

which is friendly with Washington. And you can

probably already guess that the States

have gotten busy here too. ASML

does not sell lithographs to China, so the

Celestial Empire is not yet capable of creating the most advanced chips

. Well, because there is nothing to

use. At the same time, the company's director, like

his colleague from Nvidia, is

not at all thrilled with these geopolitical games

and warns: "It

will all end the same way as with chips. China

will develop its own machines and not

only stop depending on their

lithographs, but will also begin to

seriously compete with them. But until

this happens, the fourth level and

the chains are again in the hands of the States, because the

Dutch government is dancing to their

dota. So, are the fifth and fourth

levels really going to the Americans? But

China has something to respond with here. Even the most

complex ASML lithograph can't create a chip out of thin

air. It needs materials like silicon,

gallium, germanium. All these

rare earth metals and dozens of other

components are needed, because without them,

modern electronics simply wouldn't

exist. And 70% of this market is

controlled by China, which,

in response to bans, deprives its

opponents of access to such valuable

raw materials. Because of which the States

Rare earth metals are being sought all over the world, from

Ukraine to Greenland. True, the United States

has a trump card in North

Carolina, in the town of Sprucepine, where

70% of the world's ultra-pure quartz is mined.

Such pure quartz is

truly a trump card in the hands of the United States. And

therefore, China is forced to purchase 95% of the

quartz it consumes, of course, from the

Americans from Sprucepine. But in April

26, Chinese scientists

discovered a large quartz deposit

in Tibet, which can be purified to

99.9%.

And so, having its own source of ultra-pure

quartz could free China's hands. But

extracting ultra-pure quartz is only

half the battle. Then it needs to be

made into a flint wafer. And 70% of the

global market for flint wafers for

IC chips is currently controlled by

a single Japanese company, Etsu Chemical

. This is another

bottleneck for the third Matryoshka doll. However, at the

fifth level of the raw materials chain,

both China and the United States hold strong positions.

Although the Americans, thanks to The crash of the

spheric plates 380 million years ago is

still in a more advantageous position. Does

this mean that since the United States somehow

controls every level of this

chain, the entire artificial

intelligence industry is under its complete

control? But China

disagrees and is preparing to

take over the most important point in the world

. But what kind of point is this, if the

industry has so many bottlenecks

at every stage? I'll tell you now.

Here is a map where all the points of

control over artificial intelligence are marked.

On US territory there are three flags:

first, second, and fifth level. These are

the offices of leading developers and applications

from Silicon Valley, the headquarters of the

main supplier of the entire industry

Nvidia, and the Srupai quartz mine.

At the other end of the map, on the Chinese

side, there are two flags. This is Huawei.

Well, your own Nvidia at home. And control

over raw materials, rare earth metals.

When the bird came out, the American stock

market lost a whole

trillion dollars in just one day, because no one

could have imagined that China If Deepsk develops its

model so cheaply and so quickly,

it will create competition for American

companies. However, Deepsk has only become popular

in China and Russia, while the global

market is still dominated by Cha GPT and

Gemin. That's why it has

n't yet received its own flag. Although, of course, the basic

minimum for creating a national

neural network in China has been fulfilled.

It seems that the remaining flags are scattered

all over the map. The Netherlands controls the

supply of chip production machines.

South Korea - memory production.

Taiwan - chip production. Japan -

silicon wafer manufacturing.

But all these countries are allies of the

Americans, who not only

supply the United States with everything they need, but also

impose export bans on their products

to China and friendly countries.

Of course, allies are good, but it's even

more reliable when the entire chain is on one

's own territory. Therefore, America

agreed with Taiwan's TSMC to

build US factories. And already in

October 25, the

first advanced chip was produced there. In total,

three factories will be built by 2030. However, even they are

not will cover all the needs of American

companies, so this is not a replacement, but only an

additional source for Americans to obtain the

necessary components for the industry

. [music]

Since 2020, the United States has invested

more than $645 billion in development and supply chains within the country

. These are dozens of projects

from raw material extraction and production

to research

laboratories. And today, the

Americans are far

ahead of everyone else in the world, investing a

quarter of a trillion

dollars a year in the industry. Does this mean that the United States

has won the war to create the weapons of

the future? No, the Chinese

disagree. We already know about Deep Purple and Huawei, the first

and second levels. By

2030, China plans

to import-substitute 80% of all components

needed for chip production.

Well, for example, already now 40% of domestic

demand for silicon wafers, the foundation of

any chip, is covered by its own National

Silicon Industry Group from Shanghai.

So the third matryoshka is still missing, but

the Chinese are moving towards having it

. But replacing The fourth link, the

Dutch lithographers, will be very

difficult. Right now, Chinese companies

from the EI industry are urging Xizenping

to create their own ASML. But will it work?

That's a really big question.

Without advanced chip production machines,

the Chinese are forced to make do.

They're using old lithographs and

inventing shortcuts. Well, like creating a

different architecture for the chip itself, which

allows for increased

productivity. But it's like

building a house not from modern

bricks, but from ordinary stones, but

fitting them together so that the walls

become just as strong. Well,

admittedly, it costs a lot more. The fifth

matryoshka doll only partially belongs to China

. Yes, rare earth metals

are the strength of the Celestial Empire, but quartz

has to be ordered from the Americans.

Only the launch of their own mines in Tibet and

other regions of the country will finally

free them from this dependence. And

the Europeans,

who don't want to be dependent on either the

US or China, also dream of their own chain. Europe's main strength

is, of course, the Dutch. ASML.

The fourth matryoshka isn't just in their

hands; the rest of the world is dancing with it.

The fifth matryoshka—quartz—may one day

stand alongside the fourth. In Norway,

geologists have discovered a deposit of

ultra-pure quartz. Well, it still needs to

be mined and processed. Processing,

however, won't be a problem.

The Quartz Corp. operates in Norway

, where

quartz from the American company Spruce is also processed. But the

rest of the chain is much more

complicated. Europe doesn't have its own TSMC, which manufactures

chips for [them]. It doesn't have its own Nvidia

or Huawei either. The closest equivalent of these

companies in Europe, Axelera AI, appeared in

the Netherlands only in 2021

and doesn't yet have the experience or

technology to seriously compete with the

Americans and Chinese. It turns out

that the second and third matryoshkas are out of the question, just like

the first. Currently, the most famous

neural network from Europe is the

French company Mrl AI,

which was founded in 2023.

Google alumni. Cha is the European

equivalent of GPT chat. However, even in

France itself, it accounts for less than 1% of

internet traffic. Let alone

other countries. So far, all we hear from Europe are

loud slogans about

becoming world leaders in artificial

intelligence. But how

this can be achieved is still

unclear. But the US has an answer.

Money. And right now, greenbacks are

taking over the entire AI industry

, depriving US allies of their

silicon shield. I've heard the expression "

silicon shield" or "chip shield"

when talking about your company.

Perhaps because our company

supplies a lot of chips to the world.

Maybe someone will refrain from attacking.

For Taiwan, TSMC is not just a source

of pride, but a company that

supports the island's de facto independence

. In case you didn't know, in 1949, the

Chinese government, which

had lost the civil war to the communists, fled to Taiwan.

Realizing that such a large neighbor next

door represents Faced with an existential

threat, it began to think: "How can we become

so indispensable to the entire world that we can

maintain our independence?" The choice fell

on technology. The government decided

that they didn't have enough people to become the world's factory

. Therefore, in 1973, the

Industrial Technology Research

Institute (ITRI) was established under the Ministry of Economic Affairs of Taiwan. Its mission was to

take promising technologies and

turn them into a source of income and

influence for the entire country. In

1976, ITRI

acquired a license for

microchip production technology from the Americans for $10 million. 19 of the

best engineers went to the States

to study this process firsthand. And a

year later, the first

microchip production line was launched in Taiwan. But then it

turned out that while the defect rate in America

is 50%, Taiwanese

engineers managed to reduce it to

30%. In 1985, Morris Chunk became the head of ITI

. He was born in China and

came to the United States at the age of 18, where he received an excellent

technical education at the Massachusetts

Institute of Technology. Well, and then

spent 25 years working at Texas

Instruments, a

chip manufacturer. Morris Chunk

rose through the ranks to become the

company's second-in-command, senior

vice president, because under his

leadership, the company began to

profit well from chip shipments. It was precisely this kind of

person, someone who could turn

technology into money, that was needed in

Taiwan. Chunk

was personally invited to the post of director of the institute by the

Prime Minister.

And just two years later, in 1987, Chunk

founded the very same Taiwu Semiconductor

Manufacturing Company (TSMC). Morris Chunk

proposed a simple but revolutionary

idea: to create a company that doesn't

design its own chips and doesn't

compete with customers, but only

manufactures them based on someone else's design. And while

other companies were distracted by

design, architecture, and marketing, TSMC

was able to focus on one thing: becoming the

best foundry for everyone. That's why

NVIDIA, Apple, AMD, and other

tech giants

began working with TSMC without any fear. They knew

that this foundry wouldn't steal their developments or

become a competitor. And TSMC

Meanwhile, TSMC has been perfecting its

chip production technologies for decades. And

this experience, technology, and team of

engineers cannot be copied. TSMC

has become a veritable flint shield for

Taiwan, protecting the island as well as

American missiles and guns.

However, when TSMC announced

the construction of factories on American

soil, politicians on the island

panicked that their shield had cracked.

This frightens them even more than a

Chinese attack. If a war starts, the company

will be destroyed. Everything will be destroyed.

If China takes Taiwan, it won't

get everything ready-made.

Supply chains will be disrupted, engineers will emigrate to the States,

the experience and technology accumulated over

decades will disappear. And

TSMC factories will be useless under the control of the right

party. But the blow to

China's economy without Taiwanese chips will be

terrible. China will lose 17% of its GDP,

and the rest of the world 10%. Hundreds of companies

will go bankrupt because they won't be able to

manufacture their products without chips, and

thousands of workers will end up on the street. Isn't that

so? Does anyone need this? That's a completely different

matter. If the shield simply goes to the US, then

Taiwan will be left without cover. And the

global economy won't collapse. The chips

will be churned out in American

factories. Then no one will stop

China from taking the island into its

own hands. That's how powerful

technology is today. But even the most exceptional raw materials

won't give you that kind of power.

Even a super-pure quartz deposit with

pruspine in the American outback can

be replaced somewhere in northern

Norway or mountainous Tibet. And in a

worst-case scenario, such a deposit can be

captured. No special skills are required

to dig the mineral out of the ground

. That's the real strength of

TSMC, ASML, Skyx and other [music]

bottlenecks and industries. These

companies can't be copied or

captured. They can only be offered

the fattest piece of the pie and bought out.

That's why the war of the future will not be a

war of motors or even drones, but a war

of economies, money and promises, where the winner will be

Not the strongest, but the

richest and smartest, and the world's most powerful weapon will fall into his hands

. It seems that the

US has taken the lead in the AI ​​race today. Not only have they

amassed the most important

industry players, but they're also investing enormous amounts of

money to acquire the rest.

China is still lagging behind, but it's entirely possible

that the Celestial Empire is simply not showing all

its cards. That's what happened with Deepsik, who

suddenly jumped out and said, "Boo!" "

Scared of the entire American industry."

[music] The weapons in our smartphones

are real, but while the two superpowers

are figuring out who's stronger, it seems like you and I have

nothing to fear. But as soon as

one of them gains a decisive upper hand,

the world will split into those who have

access to the cutting edge, and those who will

slide into a dependent Stone

Age without it. [music] We have a hand in determining who will win

every day, choosing

which neural network to send our request to

today: the GPT chat or Deep PSK.

So even one person here becomes a

cog in this all-powerful mechanism.

If you're afraid to miss out when

the winner of this war is clearly

determined, subscribe to this

channel, to my Telegram channel, Simple

Economy. There I explain what's

really happening in Russia and the world. This

was Nikolai Myachin with you. Bye.